The majority of the notes in this update are taken from the abbreviated TPV Developer meeting held on Friday, November 18th. The video of that meeting is embedded at the end of this update. My thanks as always to North for recording and providing it.
Server Deployments – Recap
As always, please refer to the server deployment thread for any updates.
- On Tuesday, November 15th, the Main (SLS) channel received the same server maintenance package as previously deployed to the three RC channels in week #45. This includes:
- llGetEnv() will support “region_max_prims” (feature request BUG-40825).
- llGetObjectDetails() will have a new OBJECT_GROUP_TAG function (feature request BUG-20064) – when pointed at avatars it returns the group tag you see floating above them; and also OBJECT_TEMP_ATTACHED – to tell you if something is a temp attachment (feature request BUG-5195).
- There is was planned RC deployment for Wednesday, November 16th.
No Change Window
A No Change window is in force in week #47 (week commencing Monday 21st November) in the run-up to Thanksgiving in the United States. This means that all four server channels should now remain on the same release through until November 30th, when a new RC deployment is anticipated.
[00:05] As noted in part 1 of this week’s report, the Maintenance RC viewer version 126.96.36.1991518 was promoted to de facto release status on Tuesday, November 15th. Also, and as reported in my Bento update #34, the Bento RC viewer was updated to version 188.8.131.521598. This contains three significant bug fixes, but there will be a further RC release to correct the so-called “Viagra Bug” which causes an “ever-growing groin bone”.
[02:14] Further updates are anticipated on the 360 Snapshot project viewer. The Lab is still trying to work around conflicts with the Interest List, but at the moment it still appears the best solution is to slowly pan around the area you want to photography before attempting a capture to ZIP file, although this doesn’t always take care of things.
[02:58] The 64-bit viewer builds, when they appear, will at some point include revisions to how the viewer goes about checking for updates and how the update process is managed. The aim is to make the update process more robust, and separate out the update checks from the rest of the code. There will be further discussions on this in December, which might be taken to indicate when the first 64-bit versions of the viewer might be appearing.
New Maintenance RC
[03:46] The next Maintenance viewer, containing further fixes and improvements is with QA, but should be appearing soon.
Intel HD2000 – HD4000 Graphics Issue
[58:57] As noted in my week #42 TPVD update, the Lab is going to resolve the Windows version detection issue with Intel HD2000 – HD4000 graphics on 32-bit systems by having the 32-bit version of their Windows viewer not opting into the Win 10 features, allowing 32-bit systems remain compatible with the chipsets; however, the 64-bit version will still opt-in. In addition, as the new viewer update process comes into force, the Lab is going to try to automatically update users into the correct channels, so they are no longer impacted by these issues.
Grid Status Page and Feed Changes
[04:22] The Lab is making further changes to how grid status information is displayed / delivered. This involves a move to a new status page provider, in part to provide more information over what was previously possible and to improve updates and management of grid status reports and updates.
The new grid status pages will be delivered via https://secondlife-status.statuspage.io/. Note that this isn’t actually live as yet, and any information displayed on the page is currently for testing purposes only.
The cut-over date for the new service is currently planned for Monday, November 28th, at which point a redirect will be put in place to steering people to the new status page when following dashboard links, etc. There may also be a redirect on the RSS feed URL as well. However, those wishing to use the new RSS feeds may need to manually update to one of the following URLs:
The changes are intended to allow Lab Operations to update grid status information directly from their internal chat system, rather than go through multiple steps to get the status pages updated, which is hoped will make the status information reflect updates and situations in a more timely manner, as well as the page itself having far more information available to users through the drop-down sections. There may also be a SMS feature so that those wishing to can subscribe to the service and receive grid status updates on the smartphone.
[10:18-36:22] Recent changes to the llTakeControl function, which among other things, manages interactions when in Mouselook have resulted in some users running TPVs being unable to complete Quest 3 of the new Horizons Experience as the guns for that section of the game will not function.
I’ve provided a separate report on the issue as it relates to TPVs and the Horizons Experience. However, the problem is symptomatic of wider issues with llTakeControl which also for a part of this discussion, and I refer readers the video to hear the entire conversation.
[43:58] The current voice updates are progressing through LL’s QA, but has been delayed due voice attenuation with distance and direction issues occurring on Mac systems such that anyone on the left of a listening person attenuate rapidly with distance so that at a distance of less than 10 metres, they can no longer be heard. This requires further work from Vivox to correct. Once received – which shouldn’t be too long – then the updates will be released in a project or RC viewer.
Meanwhile, the server-side Voice updates have been held pending the completion of other server-side work which is deemed to have a higher priority.
[42:36] While a fix for SL Voice crashing when plugging / unplugging a USB device was broadly fixed some time again, there are some USB devices which can still result in a Voice crash when plugged in / unplugged from a computer. These seem to be devices which aren’t correctly recognised at the time they are connected (e.g. a headset being identified as a keyboard or camera, a camera as a headset, etc.). Vivox apparently now has a fix for this issue as well which will be coming shortly.